On Tue, Feb 18, 2025 at 08:07:16PM +0000, rpinion865 wrote:
> Hey, I know how to fix that. Old story,
Another old story. Adding the 2nd 3830 disk controller to a 360/91
(and likely 3330 disk drives). Now these wern't officially supported
on the 360/91's 2860 selector channels. But by adding a "data in/out"
RPQ to the selector channel (available to allow the newer higher speed
tape drives) the 3830 disk controller can be connected.
This was the 2nd 3830 to be added, so no big deal right?
This resulted in channel control checks in the 2860.
It turns out that the 3830 can respond to the channel tags faster than
the 2860 channel can handle.
But the first 3830 is working...
Increasing the cable length of the bus and tag channel cables between
the 2860 and the 3830 fixed the problem.
PS: The 3830 really wanted a block multiplexer channel which the 360/91
didn't have. There was an IOS mod which deteced the channel
disconnect and looped until the recoonect to restart the I/O.
This didn't happen that often as there was also a program run after
each IPL which zapped the 3830 microcode so that it didn't disconnect
(a 1 bit change).
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
